|
Hello All,
?
As a possible opportunity to upgrade my? test bench's HP 333A, I recently obtained a HP 339A distortion analyzer of unknown heritage, but very late serial number.
?
Physically, not in too bad shape, all things considered:?
- missing bottom feet (not unusual, I suspect - probably was rack mounted) - I would be interested in finding these, if anyone knows of a source (new or used)
- missing end cap for input knob, otherwise all knobs complete (and not broken!)
?
- Cleaned up well, internal inspection didn't reveal any previous obvious "tinkering", looked clean. Powered up and went through basic testing (using internal oscillator), seemed to mostly work but with the following observations:
1) Output monitor completely dead (not too worried about this yet)
2) Using the internal oscillator (1V output), for frequencies below ~3.5 KHz, the Freq. Input LED (counter clockwise) illuminates. I can extinguish the LED by moving the frequency vernier knob off of the "calibrated" detent position. If I do this, then all of the lower frequencies are reported to be in the notch filter's range and autoset works under this condition, depending on the actual frequency the vernier knob has to be set anywhere from just off the detent to about 2 o'clock max. As frequencies go lower than 3.5 KHz, the vernier has to be turned more CW. I've tried various voltage output settings (3, 0.1, .03) and the above behavior doesn't change as far as I can determine. Reported distortion values seem reasonably close (if not in,) to spec.
?
What I've done so far:
- carefully deoxit only the metal contacts on all wafer switches (yes this took a while)
- re-seated the one board in a connector
- re-seated all cable and connectors
- loosened, re-tighted all mounting screws
- supplied missing endcap for input knob (my HP8640B parts bin had exactly one knob with the right cap size & color - from the output attenuator's center variable control, I believe)
?
I am going to leave the unit powered up for a couple of days, it seemed like the longer it was on, the lower I could set the frequency and not have to use the un-calibrated vernier setting (capacitors reforming??) and see where it's at.
?
- I have not compared it to my HP 333A yet, but I have hooked up a couple of other sig gens that I have just to see its operation, results seemed reasonable
- Other than the output monitor being dead, it seems like it just might need an alignment/calibration
?
I have some equipment (HP pulse generator, scopes, freq. counters, etc.), but nothing calibrated by a lab, just a semi-serious hobby for me - testing/repairing my own stuff.
I have not looked at the supply rails with a scope yet, I have looked at the output - looks like a nice sine wave.
I will review the schematic for the output monitor, but I am not hopeful if it turns out to be an active device (finding replacement transistors are pretty much unobtanium nowadays, it seems).
?
Suggestions , opinions on where to go next?
?
sorry for the TL;DR
?
Thanks,
Jim
?
I gotta say, I just love seeing the gold PCB traces on this era's equipment - no tin-wiskers here! I smile every time I open up this type of equipment.
?
|
I had this problem too, ten or so years?ago when I got my 339A.? The notch filter's lock-in range doesn't quite overlap?with the oscillator.? I don't remember exactly what I did but I think I bent the oscillator so it was centered with respect to the filter,
by tacking on caps or resistors on the troubled ranges on the A1 oscillator board.? Looking at the schematics today, I see that the A4 notch filter board has trim-pots for filter center frequency and null.? I don't know if I missed that, or was the filter
okay and the oscillator off?? Anyway, there's a place to start.
HTH,
Dave Wise
toggle quoted message
Show quoted text
Hello All,
?
As a possible opportunity to upgrade my? test bench's HP 333A, I recently obtained a HP 339A distortion analyzer of unknown heritage, but very late serial number.
?
Physically, not in too bad shape, all things considered:?
- missing bottom feet (not unusual, I suspect - probably was rack mounted) - I would be interested in finding these, if anyone knows of a source (new or used)
- missing end cap for input knob, otherwise all knobs complete (and not broken!)
?
- Cleaned up well, internal inspection didn't reveal any previous obvious "tinkering", looked clean. Powered up and went through basic testing (using internal oscillator), seemed to mostly work but with the following observations:
1) Output monitor completely dead (not too worried about this yet)
2) Using the internal oscillator (1V output), for frequencies below ~3.5 KHz, the Freq. Input LED (counter clockwise) illuminates. I can extinguish the LED by moving the frequency vernier knob off of the "calibrated" detent position. If I do this, then
all of the lower frequencies are reported to be in the notch filter's range and autoset works under this condition, depending on the actual frequency the vernier knob has to be set anywhere from just off the detent to about 2 o'clock max. As frequencies go
lower than 3.5 KHz, the vernier has to be turned more CW. I've tried various voltage output settings (3, 0.1, .03) and the above behavior doesn't change as far as I can determine. Reported distortion values seem reasonably close (if not in,) to spec.
?
What I've done so far:
- carefully deoxit only the metal contacts on all wafer switches (yes this took a while)
- re-seated the one board in a connector
- re-seated all cable and connectors
- loosened, re-tighted all mounting screws
- supplied missing endcap for input knob (my HP8640B parts bin had exactly one knob with the right cap size & color - from the output attenuator's center variable control, I believe)
?
I am going to leave the unit powered up for a couple of days, it seemed like the longer it was on, the lower I could set the frequency and not have to use the un-calibrated vernier setting (capacitors reforming??) and see where it's at.
?
- I have not compared it to my HP 333A yet, but I have hooked up a couple of other sig gens that I have just to see its operation, results seemed reasonable
- Other than the output monitor being dead, it seems like it just might need an alignment/calibration
?
I have some equipment (HP pulse generator, scopes, freq. counters, etc.), but nothing calibrated by a lab, just a semi-serious hobby for me - testing/repairing my own stuff.
I have not looked at the supply rails with a scope yet, I have looked at the output - looks like a nice sine wave.
I will review the schematic for the output monitor, but I am not hopeful if it turns out to be an active device (finding replacement transistors are pretty much unobtanium nowadays, it seems).
?
Suggestions , opinions on where to go next?
?
sorry for the TL;DR
?
Thanks,
Jim
?
I gotta say, I just love seeing the gold PCB traces on this era's equipment - no tin-wiskers here! I smile every time I open up this type of equipment.
?
|
Thanks Dave.
?
I'll look into that next. During my initial testing I did look at the freq. output of the oscillator and I didn't see anything that looked outrageous re: actual vs. setting, but I didn't run the math to see if in spec. I the cases where the input LED was on, sometimes the oscillator freq was spot on and other frequencies where the LED was on freqs. seemed off from spec, but not too far off. I think it may be a little bit of both - notch filter center freq. being off in some cases, oscillator freq. off in others.
?
I looked through the schematic at the section for the monitor output and it's a simple path from the main meter circuitry (which appears to be working) - just a resistor, fuse and then the isolation relay. I hear the relay get activated/deactivated, so once I have the unit open again I will probe that path and see where the signal stops. I am hoping that it's just the 0.062A fuse.?
?
I am going to let the unit "bake" for a week or so. After 24 hours the freq. that starts requiring the vernier is now down from 3.5 KHz to 3.3 KHz.
?
Did I read somewhere where HP did not include the schematic for the 339A's oscillator itself in the service manual?
?
JB
|
In message < [email protected]>, Jim via groups.io <james_buttons@...> writes Thanks Dave.
?
I'll look into that next. During my initial testing I did look at
the freq. output of the oscillator and I didn't see anything that
looked outrageous re: actual vs. setting, but I didn't run the math
to see if in spec. I the cases where the input LED was on,
sometimes the oscillator freq was spot on and other frequencies
where the LED was on freqs. seemed off from spec, but not too far
off. I think it may be a little bit of both - notch filter center
freq. being off in some cases, oscillator freq. off in others.
?
I looked through the schematic at the section for the monitor
output and it's a simple path from the main meter circuitry (which
appears to be working) - just a resistor, fuse and then the
isolation relay. I hear the relay get activated/deactivated, so
once I have the unit open again I will probe that path and see
where the signal stops. I am hoping that it's just the 0.062A fuse.
?
?
I am going to let the unit "bake" for a week or so. After 24 hours
the freq. that starts requiring the vernier is now down from 3.5
KHz to 3.3 KHz.
?
Did I read somewhere where HP did not include the schematic for the
339A's oscillator itself in the service manual?
? I think it is there are some versions for this unit online here and there had one for years very reliable;! Whilst your at it make sure of you change the power supply caps this improved the distortion reading on mine by a decent amount supply ripple i suppose?>. -- Tony Sayer
|
On Sun, Sep 8, 2024 at 08:40 AM, Jim wrote:
missing bottom feet
Yes, they often walk away here... hide amongst a large number of friends... in an un-labeled? cardboard box somewhere... and eventually...enjoy the ride... to their final home... the dump.
There are models for them... and they can be 3D printed...by the local teenager... and if all you are interested is in functionality... from a distance... they look reasonably similar.
The bright orange PETG print threw me off at first... but, one can't see it when the unit is on the bench... and the 'feet' keep the case from contacting the bench... which is a 'no no' here.
But of course... nothing beats the bone fides.
You can always find those on Ebay, for 10.00 to 20.00 USD + plus 50.00 USD shipping.
I think scrappers, junk magnets, and metal hounds, keep the feet, and 'throw' the instrument away.
?
|
On 9/9/24 12:53, Roy Thistle wrote: missing bottom feet
Yes, they often walk away here... hide amongst a large number of friends... in an un-labeled? cardboard box somewhere... and eventually...enjoy the ride... to their final home... the dump. Wow, Roy. -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
Re feet there is a file for feet to make them using a 3 D printer ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Roy Thistle
Sent: 09 September 2024 17:53
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] HP 339A Distortion Test Set repair/refurbish (hamfest find) (long read) ? On Sun, Sep 8, 2024 at 08:40 AM, Jim wrote: missing bottom feet
Yes, they often walk away here... hide amongst a large number of friends... in an un-labeled? cardboard box somewhere... and eventually...enjoy the ride... to their final home... the dump. There are models for them... and they can be 3D printed...by the local teenager... and if all you are interested is in functionality... from a distance... they look reasonably similar. The bright orange PETG print threw me off at first... but, one can't see it when the unit is on the bench... and the 'feet' keep the case from contacting the bench... which is a 'no no' here. But of course... nothing beats the bone fides. You can always find those on Ebay, for 10.00 to 20.00 USD + plus 50.00 USD shipping. I think scrappers, junk magnets, and metal hounds, keep the feet, and 'throw' the instrument away.
|
Hi Tony what Distortion level did you manage to get the 339 down to ? also
what is the level flatness you have managed to achieve between say 75 hz to
20 Khz ?
toggle quoted message
Show quoted text
-----Original Message----- From: [email protected][mailto: [email protected]] On Behalf Of tony sayer Sent: 09 September 2024 11:28 To: [email protected]Subject: Re: [HP-Agilent-Keysight-equipment] HP 339A Distortion Test Set repair/refurbish (hamfest find) (long read) In message < [email protected]>, Jim via groups.io <james_buttons@...> writes Thanks Dave.
?
I'll look into that next. During my initial testing I did look at
the freq. output of the oscillator and I didn't see anything that
looked outrageous re: actual vs. setting, but I didn't run the math
to see if in spec. I the cases where the input LED was on,
sometimes the oscillator freq was spot on and other frequencies
where the LED was on freqs. seemed off from spec, but not too far
off. I think it may be a little bit of both - notch filter center
freq. being off in some cases, oscillator freq. off in others.
?
I looked through the schematic at the section for the monitor
output and it's a simple path from the main meter circuitry (which
appears to be working) - just a resistor, fuse and then the
isolation relay. I hear the relay get activated/deactivated, so
once I have the unit open again I will probe that path and see
where the signal stops. I am hoping that it's just the 0.062A fuse.
?
?
I am going to let the unit "bake" for a week or so. After 24 hours
the freq. that starts requiring the vernier is now down from 3.5
KHz to 3.3 KHz.
?
Did I read somewhere where HP did not include the schematic for the
339A's oscillator itself in the service manual?
? I think it is there are some versions for this unit online here and there had one for years very reliable;! Whilst your at it make sure of you change the power supply caps this improved the distortion reading on mine by a decent amount supply ripple i suppose?>. -- Tony Sayer
|
Great find. The 339A is a great improvement over the 333. I have the 339A and 334A, both are very useful instruments. Mine has been seeing a lot of work lately with my current project. Let me know if you need any baseline measurements on the 339A and I¡¯ll try and help you out. Once you get it working I¡¯d recommend checking online for the improvements which can be made. There¡¯s a very comprehensive list on one of the audio forums (audio karma I think). Mine bottoms out around 0.0012% at 1khz but that can be improved.?
|
I have a 339A that I am about to press into service unfortunately it hasn¡¯t been switched on for over 5 years so will probably require a bit of attention I have uploaded the service manuals to this sight Could I ask if you find any information or notes that you upload them to the files on this sight I am also interested in the baseline measurements on the 339A you talk about Regards Paul ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Jeremy B
Sent: 09 September 2024 20:13
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] HP 339A Distortion Test Set repair/refurbish (hamfest find) (long read) ? Great find. The 339A is a great improvement over the 333. I have the 339A and 334A, both are very useful instruments. Mine has been seeing a lot of work lately with my current project. Let me know if you need any baseline measurements on the 339A and I¡¯ll try and help you out. Once you get it working I¡¯d recommend checking online for the improvements which can be made. There¡¯s a very comprehensive list on one of the audio forums (audio karma I think). Mine bottoms out around 0.0012% at 1khz but that can be improved.?
|
Hi Jim, et al.
I have an HP339A on my bench so I'm following this discussion thread with interest. My unit was manufactured in 1983 and currently needs some TLC. I offer some of the things that I have learned as I used my unit.
?
On Sun, Sep 8, 2024 at 06:54 PM, Jim wrote:
...I looked through the schematic at the section for the monitor output and it's a simple path from the main meter circuitry (which appears to be working) - just a resistor, fuse and then the isolation relay. I hear the relay get activated/deactivated, so once I have the unit open again I will probe that path and see where the signal stops. I am hoping that it's just the 0.062A fuse.?
Note that the monitor output is disabled when the OSCillator LEVEL FUNCTION is selected. Reference manual sections 3-28 & 3-37. I did need to replace the 62mA fuse in my unit after I inadvertently applied an external signal to the front panel Output Monitor terminals. I was very happy to learn that the instrument was protected from such operator mistakes.
On Sun, Sep 8, 2024 at 06:54 PM, Jim wrote:
Did I read somewhere where HP did not include the schematic for the 339A's oscillator itself in the service manual?
My manual does include a foldout schematic of the oscillator circuits. However, the oscillator circuit uses two OP Amp ICs that are listed as manufactured by HP.
?
When I was considering the purchase of my unit it was still easy to talk to a live person at an HP service center. The rep told me that "The most common failure in the 339 is the switches which wear out or get dirty. The frequency switches wear out the most with the distortion range switch next. It takes about 2 hours to change the switches so they often just change the complete PCB assembly. Both approaches are expensive. They don¡¯t use any special cleaner just a good quality contact cleaner."
?
Enjoy your 339A Jim.
Cheers,
Tom.
|
The oscillator IC is the Harris HA2625 which is discontinued. I believe the LT1468 is a suitable replacement. The HA2625 is also used as the product amplifier to drive the true RMS converter IC, an AD536 (still available albeit expensive). The HA2625 is used elsewhere in the unit, once you know one HP part number you can find all the similar ICs. The AD536 is a great chip for RMS if you drive it with a large signal. It¡¯s very accurate below about 200-300khz. The AD637 is better if you want more frequency response but the pins are completely different (I digress). The oscillator uses a VCR2N fet for level control. You can find them second hand or try a 2N4091 or 2N4092 which are available from Mouser. The 4091 being the better choice.?
|
The is also a modification for a second harmonic trim which uses feedback around the fet via a 2k trimmer. I have implemented this in a clone I built of the oscillator section and it makes a very marked improvement in reducing the 2H level.?
|
In message < [email protected]>, Paul Bicknell <admin@...> writes Hi Tony what Distortion level did you manage to get the 339 down to ? also
what is the level flatness you have managed to achieve between say 75 hz to
20 Khz ?
It was just that Flat!, and IIRC it was around .001% some of course was just noise.. -----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of tony sayer
Sent: 09 September 2024 11:28
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] HP 339A Distortion Test Set
repair/refurbish (hamfest find) (long read)
In message <[email protected]>, Jim via groups.io
<james_buttons@...> writes
Thanks Dave.
?
I'll look into that next. During my initial testing I did look at
the freq. output of the oscillator and I didn't see anything that
looked outrageous re: actual vs. setting, but I didn't run the math
to see if in spec. I the cases where the input LED was on,
sometimes the oscillator freq was spot on and other frequencies
where the LED was on freqs. seemed off from spec, but not too far
off. I think it may be a little bit of both - notch filter center
freq. being off in some cases, oscillator freq. off in others.
?
I looked through the schematic at the section for the monitor
output and it's a simple path from the main meter circuitry (which
appears to be working) - just a resistor, fuse and then the
isolation relay. I hear the relay get activated/deactivated, so
once I have the unit open again I will probe that path and see
where the signal stops. I am hoping that it's just the 0.062A fuse.
?
?
I am going to let the unit "bake" for a week or so. After 24 hours
the freq. that starts requiring the vernier is now down from 3.5
KHz to 3.3 KHz.
?
Did I read somewhere where HP did not include the schematic for the
339A's oscillator itself in the service manual?
? I think it is there are some versions for this unit online here and there
had one for years very reliable;!
Whilst your at it make sure of you change the power supply caps this
improved the distortion reading on mine by a decent amount supply ripple i
suppose?>.
--
Tony Sayer
-- Tony Sayer Bancom Communications U.K. Tel+44 1223 566577 Mob: 07707799230 4 Wingate close, Cambridge, England, CB2 9HW E-Mail tony@...
|
Dave McGuire
Paul Bicknell
Tom Phillips